The false dichotomy: OSM as open or closed

I want to nip something in the bud before it gets too sexy to repeat by the geo-intelligensia who’re being herded in to place by the jetset geospeaking crew.

Here’s the basic argument: “openstreetmap is closed because the ordnance survey released data under a ‘more open’ license.” there are a bunch of obvious ways of rephrasing it.

It sounds simple, it’s a nice meme that will take hold with an audience. It’s one of those “Oh wow! A deep and logical point I haven’t thought of before” kind of mind bombs you get in a Malcolm Gladwell book. You can grin and snicker with your friends at the pub: Those silly OSMers! They aren’t even open. They’ve been leapfrogged by good old central government planning! If only they were _truely_ open, then we could all use the data!

Let’s strip this apart.

First, it’s a false dichotomy. That means they’re reducing OSM to one of two states namely ‘open’ and ‘closed’ when in fact there is a spectrum in between. There are many, many open source licenses which fan out between non-commercial with some sharing rights, right through to the public domain which allows you to do virtually anything you like. Using extremely loaded terms like ‘open’ and ‘closed’ is a false dichotomy. It’s a neat hand wavy argument designed to make the speaker look intelligent when comparing, actually, apples and oranges which we’ll get on to in a second. Everyone has a happy place on the license spectrum and this isn’t a point about where you the reader lay on that line, but a point about framing the argument in to two unrelated positions to make it look simple when it isn’t. And then be witty by positioning OSM on the closed end. Which it’s not.

Next let’s look at the type of data we’re comparing. This gives them some wiggle room to say “oh we didn’t mean the OS data but this other improbable dataset over there”. But let’s stick with the wonderful Ordnance Survey data release. You can’t easily compare the OS data release and OSM data for multiple reasons.

One is that OSM is a moving target. The OSM data is improving continually from the work of a large and diverse community behind it. The OS data is not. It is improving, granted based on a lot of employee surveyors, but there’s no guarantee (correct me if I’m wrong) that you will still have open OS data up to date in 2015. OSM has that guarantee. It’s in the license. We can’t fork OSM in to a proprietary stack. You (if you’re the OS) can fork the OS data, and of course they very reasonably do. They have a copy they can license to whoever they like under whatever terms, and also the open copy. Those aren’t guaranteed to be the same thing in perpetuity. The free version might start to lag the proprietary one. The free one might cease to be released. You have no idea, and no control over that. With OSM you have every control and every assurance that it will always be open.

Here’s some more simple kickbacks – OSM is global not just in the UK. OSM maps more types of things than the OS maps. OSM has a community around it of tools and software and mappers. Stop and think for a second – would you get all of those things under a ‘more open’ license? The answer is not very clear and I suspect the answer is no, but everyone has an opinion on that.

So the datasets are different. You can’t compare apples with oranges, the GDP of a country with the market cap of a company, or a free and open global community continually improving data with a point release of a single dataset in one nation.

Now dear reader, allow me to make my own cheap shot. The very people dancing around with this argument tend to work for large companies…. with large proprietary datasets. Perhaps Google could release the MapMaker data? Or perhaps Nokia could release their NavTeq data? Ghandi wanted us to become the change we wanted to see in the world. That means beginning at home. Surely we should be hearing more from the valiant geovisionaries during their globetrotting speaking tours (speaking as a globetrotting speaker) about their fine efforts at home to release large chunks of data? Say under an ‘open’ license like the OS did? Sadly I find silence on this issue, but it’s not a surprise. Oh and putting on your blogs and talks that everything is your personal opinion doesn’t make all these arguments disappear guys.

Richard Stallman has many failings as we all do, but one thing he repeats resonates with me on OSM – don’t give up the long term vision for some short-term gains. To me that means not moving to public domain because one company might donate some data to us, or staying with another license because another will. There are other, valid, arguments for those moves but a short term gain is not. These are admittedly very attractive offers. Imagine if we moved to license X and Google gave us a lot of aerial imagery, or license Y and Nokia delivered us to the promised land.

But wait a minute, if they believed in those things then surely they’d gift them to us anyway? That’s rarely the case. And I applaud those who do, like Yahoo and AND for example.

What you’re seeing is that OSM is somewhere like one third or one half way toward real maturity. Time is very skewed. In some ways the 6 years we’ve existed looks like a very long time to those involved. If only we could get things quicker by changing the fundamental structure of OSM so that they would, perhaps, grant us some data (for that is often the offer)? On the other hand, 6 years compared to the timespan the OS took to map the UK, or any company or agency took to map a dataset as large as OSM… is tiny. We’re talking 6% of the time in some cases. Yes, it can take 100 years to map. Many datasets are built quicker, but very very few on the order of 6 years. Globally. For free. Then released for free.

As we’re in this midpoint somewhere and the pillars of proprietary data begin to collapse around us you see panic. You see people propping them up with beams. You see new temples built of wood made to look like stone. You see a general shoring up of an obsolete structure and in the general thrashing around you see a lot of people looking for someone to blame while they hold on until a secure future arrives.

And who shall we blame and poke? Google? Inscrutable. Nokia? They make phones.

No, OpenStreetMap will be kicked and get the blame. We’re the easiest target, with the most to lose. As far as I can see the cheap ‘OSM isn’t open’ arguments are just a way to deflect attention from the pillars falling: Think not of us releasing our vast datasets, think instead of OSM not being open. Think not of us contributing, think instead of changing OSM and allowing us to use it as we see fit – perhaps then we will contribute. Think of us as not evil, think of OSM as splintered, hemorrhaging and not as united as us.

If you have some of that OSM vision then you should be thinking about where OSM will be in another 6 years, not what we have to give up to appease those who’d love our data without attribution or reciprocal rights. It will be okay. Relax. In a few years it will look silly that anyone ever thought we should give up our values for some quick data injections.

OpenStreetMap has a challenge rebutting all these arguments. It’s tiring. We’re nice people. We have better things to do. We speak with many voices, each with a different view on key things like licensing, whereas those guys have voices almost in choral harmony. But don’t mistake that quiet for some kind of general acceptance. We haven’t been sleeping while you suddenly decided OSM was ‘closed’.

So, dear reader, the next time you see a talk or blog post where OSM is darkly described as ‘closed’ feel the strength to ask if the speaker works for a company with a vast proprietary dataset and if they’d like to discuss opening it, or perhaps if they have any experience whatsoever in open projects, licensing or building communities. If they mutter about their legal team not being comfortable with OSMs license, you can invite them to look at all the other ginormous companies using OSM who are perfectly comfortable.

Oh, and say hi from me!